Light Sheet Sectioning

01:00:11.14 Hello, my name is Ernst Stelzer. I'm from the Goethe
01:00:15.19 University in Frankfurt am Main in Germany.
01:00:18.11 And I will introduce you to light sheet based fluorescence
01:00:21.06 microscopy. And in the course of this talk, I'll tell you why
01:00:26.01 such a system reduces photobleaching and phototoxicity
01:00:28.19 in three-dimensional imaging, why this is one of the
01:00:32.24 technologies of the future. Before I start, I would like to give you
01:00:37.18 an idea of who worked on this talk, and who contributed
01:00:41.21 to it. So, Alexander Atzberger is a person who did a lot on the
01:00:46.10 instrumentation side. A lot of the biological data actually comes
01:00:51.02 from Francesco and Nari, and Daniel von Wangenheim, and
01:00:55.11 Christian Mattheyer contributed some of the videos that you
01:00:58.15 are going to see in the next 30 minutes. Some people who worked
01:01:04.05 in my group during the past few years and who also contributed
01:01:07.00 quite a lot are Fabian Harle, who worked on the hardware,
01:01:11.00 Philipp Keller, and Uros, Jan Huisken, who were PhD students
01:01:17.03 in my lab. And also Klaus Greger and Jim Swoger, who
01:01:21.14 contributed to the technology development. And of course,
01:01:25.19 Emmanuel Reynaud, who helped us with establishing
01:01:29.12 the sample preparation technology. So, I will introduce
01:01:35.11 you to light sheet based fluorescence microscopy. I'll talk about
01:01:38.04 optical sectioning, I'll give you an idea of what multiple-view
01:01:41.16 imaging is, and as I said before, I'll talk about reductions in
01:01:47.03 photobleaching and phototoxicity. I'll also talk a bit about
01:01:51.19 the applications, so I'll give you an idea why it's important
01:01:55.14 in development biology and also in plant biology. And at the
01:01:59.05 very end, I'll show you what current systems look like.
01:02:03.00 So, over the past ten or twelve years, we worked on the physics,
01:02:09.11 as well as on the instrumentation, we applied the instruments
01:02:14.06 in developmental biology and in cell biology, as well as in biophysics.
01:02:18.07 And we also developed some of the software that is necessary
01:02:21.03 to actually analyze the huge amounts of data that I generated
01:02:27.03 with such a system. So the motivation for us is that we want to work
01:02:32.00 in the life sciences. So we want to work with live specimens,
01:02:35.18 they should be three-dimensional, we also want to record them in
01:02:39.21 three dimensions. And we want to look at multiple processes
01:02:43.03 as a function of time. And of course, we want to monitor
01:02:46.17 these processes with a certain precision or resolution over a
01:02:52.14 certain period of time. And these periods of time can be very long,
01:02:55.17 and they can also be very short. And most appropriate,
01:03:01.17 spectral techniques that are spatially and temporally
01:03:06.18 resolved, such as lifetime imaging, or just fluorescence microscopy.
01:03:12.07 And of course, we want to take advantage of fusion proteins,
01:03:15.13 and the physiological context also has to be maintained.
01:03:19.11 Now, a big issue that you always have to consider when you work
01:03:26.16 with fluorescence microscopy is that the number of fluorophores
01:03:29.08 in a specimen is finite. So when you look at a specimen,
01:03:33.06 or at a section of a specimen, you have to be aware of the fact
01:03:38.01 that at any given time for this section, the number of fluorophores is
01:03:43.07 finite. And since the fluorophores are consumed during the observation
01:03:48.08 process, also the number of photons is finite. And therefore,
01:03:54.14 the three most important topics in fluorescence microscopy
01:03:57.04 are photons, photons, and photons. It's all about photons.
01:04:02.07 And there are various different ways you can actually
01:04:07.09 change the optical system, you can vary that a little bit
01:04:13.13 where you get your photons from, but in the end, it turns out
01:04:16.11 that you can only vary the object size. You can change the
01:04:19.14 resolution, you can play a little bit with the signal to noise
01:04:22.20 ratio, and you can also change the time period that you observe.
01:04:25.21 But essentially, it's the product of those four terms that determine
01:04:30.12 the number of photons that you can get out of a specimen.
01:04:34.17 And you can change the resolution, but that will affect the
01:04:38.17 time period over which you can observe a specimen.
01:04:42.10 You can change the object size, but that will again affect
01:04:45.20 the signal to noise ratio, etc ... So, photobleaching is the big
01:04:53.02 topic in microscopy. And if you have a regular epifluorescence
01:05:00.00 microscope or system that uses the same lens for illumination
01:05:03.02 detection, what happens usually is that the light passes
01:05:06.17 more or less through the specimen. So if you could place a detector
01:05:12.02 anywhere within your specimen, you would measure the same
01:05:15.06 laser or excitation intensity in each of those planes.
01:05:20.04 And that of course means that as long as you stick to the
01:05:23.23 linear regimen, you will also excite the same number of
01:05:28.23 fluorophores in each of those planes. And if you continue this line of
01:05:33.20 thought, and look at what happens when you record an entire
01:05:37.11 plane, you will realize that wherever you record such a plane,
01:05:42.12 you actually excite all the fluorophores in the entire specimen.
01:05:47.18 So if you have a specimen that requires about twenty
01:05:51.09 different planes, by the time you've reached the tenth plane,
01:05:55.10 you've already excited all the fluorophores in the entire
01:05:59.22 specimen nine times. So you've already wasted about 90%
01:06:04.06 of your signal. Another important aspect is the intensity level at which
01:06:12.14 you work. So, if you look at the solar constant, which is about
01:06:17.07 1.4 kiloWatts per square meter at the equator, or about 1 kiloWatt
01:06:23.14 per square meter in central Europe, you get a feel for
01:06:31.05 how much energy we are actually exposed to in a natural
01:06:38.18 environment. Now on a microscopic level, this boils down to
01:06:43.07 about 1 nanoWatt per square micrometer, or about 100 milliWatt
01:06:46.13 per square centimeter. And this is not a lot. And if you
01:06:54.03 expose yourself somewhere in the sun, let's say for about
01:06:57.05 ten minutes, your skin will already start to turn red if you're
01:07:04.06 a caucasian. So, your body already responds to these low
01:07:10.22 energy levels. And from that, you can derive, let's say the
01:07:16.15 energy level to which you can expose a cell or an embryo.
01:07:19.20 I'm not trying to tell you that these are exactly the levels
01:07:22.24 that biological systems should be exposed to or the levels
01:07:28.06 that should not be exceeded, but at least it gives an idea
01:07:31.06 for the range to which we can, or the levels to which we can
01:07:36.03 expose biological specimens. And we have to be very careful
01:07:40.05 not to exceed them too far. And if you look at the kind of
01:07:44.06 microscopy that you are using, you'll probably notice
01:07:46.15 that these values in your instrument for most of your experiments
01:07:50.15 are probably much, much higher. Now, let's have a look
01:07:58.01 at the design of these light sheet based microscopes.
01:08:02.08 And we'll start by having a look at some of the developments
01:08:06.09 that come from the past, particularly the Theta microscope.
01:08:08.20 So, instead of using an epifluorescence microscope, which uses
01:08:15.11 the same lens for illumination and detection, we actually
01:08:19.00 split the illumination and detecting into two different
01:08:24.08 systems. So we have one lens which we use for the
01:08:27.05 illumination, and the second lens which we use for the
01:08:30.18 detection. And these lenses are arranged at an angle
01:08:33.22 of 90 degrees, and that's why we call this a Theta microscope.
01:08:36.18 And when we start with Theta microscopy, we actually had a point
01:08:42.11 illumination and a point detection, which is just sketched here
01:08:46.01 with the simplified point spread functions. Which overlap
01:08:50.16 in the region that you are actually interested in.
01:08:53.18 And in these systems, what you have is you illuminate
01:08:58.17 volume elements that you never observe, and on the other hand,
01:09:01.15 you observe volume elements which have never seen any light.
01:09:04.14 So in essence, you take the product of those two point
01:09:09.17 spread functions, and that gives you an idea of the volume
01:09:12.17 that you're actually looking at. And this is something we could demonstrate
01:09:16.17 here with simple latex beads. So you see here, for example,
01:09:20.23 if you take this Theta microscope, you get a really nice
01:09:26.16 round latex beads, whereas with a confocal fluorescence
01:09:29.10 microscope, you get this extension along the optical axis. That
01:09:33.16 works for small, as well as for large objects. Now, you can continue
01:09:39.06 this. And in the light sheet microscopy, we don't use a high NA
01:09:44.10 illumination, but we rather use a low NA illumination.
01:09:46.20 And we also do not illuminate just a single point in the specimen, but
01:09:51.10 we rather illuminate an entire plane. And therefore, we also
01:09:55.21 don't use a point detection device, but rather a camera,
01:10:00.11 which allows us to record multiple or several million pixels
01:10:05.04 at the same time. But what's still the case is that we illuminate
01:10:12.10 parts of the specimen, or that we observe parts of the specimen
01:10:15.23 rather, that we have never illuminated. This is the real big distinction
01:10:21.01 or the big advantage of such a system. So essentially,
01:10:25.16 we have this simple light sheet, and this light sheet overlaps
01:10:29.13 with a focal plane of the detection system. But in the light sheet
01:10:34.22 based microscopes, there's no light in front and there's no
01:10:38.09 light behind the specimen. And therefore, the parts that are
01:10:43.21 outside of the light sheet cannot contribute to the image. So there is no
01:10:47.21 out of focus blur. And in addition, since we do not illuminate those
01:10:52.05 parts in front and behind the light sheet, we also do not induce
01:10:57.03 any photobleaching and no phototoxic effects in these specimen.
01:11:01.13 And then of course, we can move the specimen through the light sheet
01:11:07.12 or the light sheet through the specimen, and in this way we can get
01:11:10.20 three dimensional information of the specimen. So here you see
01:11:16.05 stacks that were recorded. Slices through a specimen.
01:11:20.22 And then also the reconstruction of those objects in a computer.
01:11:25.07 And up there, you see a copepod, here is a polycystic mouse kidney.
01:11:29.24 Down there are parts of a zebrafish, and over here is a
01:11:34.08 part of an arabidopsis plant. And you will notice that
01:11:39.21 the images are very crisp, very much like what we
01:11:42.20 expect from a confocal fluorescence microscope, but
01:11:45.24 the big difference here is that the recordings are much, much
01:11:49.16 faster. So we record here probably around 10 frames per second.
01:11:53.09 The dynamic range is much higher, so it's not only 6 bits, but
01:11:58.06 rather more like 10-12 bits. And the number of pixels is much larger,
01:12:04.13 so you can easily record anywhere between 3-5 million pixels
01:12:10.20 in a single frame. Now, another thing that you can do, you don't have
01:12:19.05 to, but it makes a lot of sense, is you can change the optical
01:12:23.02 arrangement. So instead of using an upright or an inverted
01:12:25.23 microscope, you can place the lenses essentially into the
01:12:30.10 same plane in which you mount this microscope.
01:12:32.21 And then the specimen is inserted along an axis that is parallel
01:12:37.16 to gravity. And that means we can rotate the specimen
01:12:41.22 without actually affecting the forces that apply to the specimen, so
01:12:46.13 the specimen is not distorted as we rotate it. We usually also
01:12:54.04 place everything inside a such a chamber, which can be filled with a
01:13:00.00 liquid. We can also use water dipping lenses, we can also use this for
01:13:04.06 temperature control and various other means. Now,
01:13:13.04 as I said, we use this light sheet and we can record stacks
01:13:17.15 of images, but due to the fact that we have this special
01:13:20.16 arrangement of the system, the holding of the specimen along an axis
01:13:27.18 that is parallel to gravity, we can rotate the specimen. And in this way,
01:13:31.20 we can record data sets along different directions. And this
01:13:36.20 gives us kind of the complementary views of the specimen.
01:13:39.11 This is really important in microscopy because a microscope
01:13:45.12 or a light microscope, usually has a much worse axial resolution
01:13:48.19 than lateral resolution. And I can demonstrate this quite easily
01:13:53.01 if I just hold my hand in front of my face, you'll have a
01:13:57.15 very hard time even if this were an audience and you could really see
01:14:01.14 everything with two eyes, to actually estimate the distance of the hand
01:14:05.12 from my face. However, if I rotate myself by 90 degrees,
01:14:10.12 you can very easily measure this distance. And that is essentially what
01:14:15.24 we do in a computer, we take those different data sets, and
01:14:19.10 then we replace the bad axial resolution by the good lateral
01:14:23.12 resolution. And then we get data sets such as this.
01:14:27.21 Up here, you see an example, that is a yeast cell which was
01:14:31.14 recorded with a light sheet along a single direction. And
01:14:34.21 here you see a combination of various views that Uros actually
01:14:40.14 produced. Down here, you see a data set from Jim Swoger
01:14:43.10 which he recorded along a single direction. It's a drosophila
01:14:48.05 embryo. And then he fused eight different data sets, which
01:14:51.23 he recorded along eight different directions. And you notice
01:14:55.14 you don't see all the cells here, but here you see all the cells
01:14:58.15 and in addition, you have a much better resolution.
01:15:01.14 That you can do for small objects, such as the yeast up here.
01:15:05.05 You can work with cysts, you can work with medaka fish.
01:15:10.14 Zebrafish embryos. Copepods. Asters. Spheroids. And
01:15:16.05 also with plant cells or whatever is necessary for your
01:15:21.22 kind of experiments. It can work with live, as well as with
01:15:24.21 fixed specimens. Now the big issue, as I said, is the confined illumination.
01:15:32.08 And we can quantify that relatively simply. So here we have
01:15:39.17 a medaka fish juvenile, where we just see the head. And this is a
01:15:45.24 relatively thick specimen, several hundred microns. And if
01:15:49.14 we now illuminate this with a regular microscope, as I said
01:15:52.23 before, we always illuminate the entire specimen. However, with the
01:15:56.20 light sheet, we only illuminate a fraction of that specimen. And this
01:16:03.20 fraction is essentially determined by the thickness of the light sheet.
01:16:06.04 And if we now take the ratio of the size of the object, divided
01:16:11.00 by the thickness of the light sheet, we get a number that
01:16:17.16 varies quite dramatically with the size of the specimen.
01:16:21.14 So, here in my opinion, it's somewhere between 3 and 1000.
01:16:26.06 Let's have a look in more detail. So if we have a yeast cell,
01:16:30.21 which is relatively small, we get an improvement which is somewhere
01:16:34.08 between 3-6. However, if we have a zebrafish, which is much larger,
01:16:38.15 the improvement is already a factor of 200-300. So what
01:16:42.16 does this number mean? It means that if you take regular fluorescence
01:16:47.20 microscope and try to record a stack of images of a zebrafish,
01:16:52.07 you require 200-300 times more light or you expose the specimen
01:16:57.13 to 200-300 times more light than if you use the light sheet based
01:17:02.03 fluorescence microscope. So you're exposing it to 2-3 orders of
01:17:06.12 magnitude more light, which is of course, quite dramatic.
01:17:09.12 And if you now use the confocal fluorescence microscope,
01:17:13.17 this number would increase to a factor of 5000-6000.
01:17:18.15 And that of course means that with a light sheet based
01:17:21.18 microscope, you can perform totally different experiments
01:17:26.01 than you were ever able to do with regular and confocal
01:17:30.03 fluorescence microscopy. And that is also the reason why
01:17:36.01 these systems have become relatively popular now for
01:17:39.07 developmental biology. Here's an example from our collaboration with
01:17:46.04 Johann Wittbrodt and his PhD student, Annette Schmidt.
01:17:51.04 And Annette actually injected H2B-eGFP mRNA into the very
01:17:56.16 early zebrafish embryos, and then Philipp took those
01:18:01.15 specimens, put them into the microscope and observed
01:18:03.23 them for up to three days. And what you see here is
01:18:07.04 essentially just the first day of this development. And he recorded
01:18:11.10 about 200 stacks every 90 seconds, from the front, as well as
01:18:18.00 from the back of the specimen. So this is exactly the same specimen,
01:18:22.01 and essentially, you see all the data or all the cells during
01:18:26.14 this first period. And since the quality of those data sets is
01:18:31.01 very, very good, you have a very good signal to noise ratio
01:18:33.14 here, and you can easily distinguish the different cells in the
01:18:37.21 original data. He was able to track these cells and get
01:18:45.03 an unbiased view of the direction along which those cells
01:18:48.17 move. And in green, you see an upward movement. In cyan,
01:18:51.23 you see a downward movement. In yellow, you will see movements
01:18:55.06 to the left and to the right. And this gives an idea of the complexity
01:19:01.03 of the movement of the cells during the early phases here
01:19:06.18 of zebrafish development. Now, the next slide just summarizes some
01:19:15.05 of those pieces of work. Again, you see the two data sets here
01:19:19.00 in this reconstructed view. Up here, you see views of
01:19:24.15 a drosophila embryo along four different directions,
01:19:28.21 also recorded in the same way as this system.
01:19:32.00 Here you see a zebrafish during the third day of its
01:19:37.09 development. And down here, we have more specialized
01:19:40.07 views, where we've picked parts of the data of such data sets
01:19:44.07 to look, for example, at eye development. But of course,
01:19:49.15 this can work with any specimen that you can label and where you're
01:19:54.02 really interested in long term evolution.
01:20:00.20 Now, light sheet based fluorescence microscopy is what I regard as
01:20:05.10 a kind of technology which has various different implementations.
01:20:09.18 So we've implemented it as a single plane illumination
01:20:12.13 microscope, or SPIM. Then we've also implemented it as a
01:20:16.04 digital version, which we call DSLM, which essentially creates
01:20:19.17 a light sheet with a scanner, which we can move up and down.
01:20:22.10 And then on top of that, we implemented incoherent
01:20:26.11 structured illumination, a laser based nanoscalpel, fluorescence
01:20:33.06 correlation spectroscopy, FCS, and of course, also fluorescence
01:20:37.01 lifetime imaging, and FRET. This is possible because the
01:20:42.16 detection part is essentially just a regular fluorescence
01:20:46.01 microscope. And whatever can be done with a fluorescence
01:20:48.10 microscope, you can probably also do with one of those light sheet
01:20:52.13 based fluorescence microscopes. Now other groups have
01:20:56.03 also done some really excellent work. There are several groups who
01:21:00.16 concentrate on clearing, they usually refer to their microscopes as
01:21:05.15 ultramicroscopes. There are people who introduced two
01:21:11.00 photon microscopy to improve the depth penetration, and also
01:21:15.17 annular illumination or Bessel beams, to improve the resolution
01:21:21.19 of those systems. And then there are more specialized versions, such
01:21:24.24 as OCPI or iSPIM, which are used for the observation of
01:21:31.12 calcium waves, or of C. elegans embryos. And there are probably many more
01:21:38.02 around and outside that are used for all sorts of applications.
01:21:43.15 Now, what we've done for developmental biology, which has worked
01:21:49.18 out really well, is something that we're not trying to push for
01:21:53.09 plant biology. Now, the system of choice in plant biology is obviously
01:21:58.23 Arabidopsis thaliana. So, you can look at basically all the
01:22:03.24 parts of this plant. But here, we will concentrate on the
01:22:09.15 root tip and on the lateral root development, in particular.
01:22:13.23 The nice thing is you can, in the light sheet based microscope,
01:22:17.24 embed the specimen in exactly the same way as you
01:22:23.12 place it into the culture chamber. So you can grow them into these gel
01:22:28.16 layers, and then you just put them into the light sheet
01:22:32.08 based microscope. And this is summarized here in this
01:22:36.18 sketch. So, what we see is that the plant grows in an upright
01:22:41.12 position, which is natural. The root is in the medium, the leaves
01:22:47.17 remain in the air. We can have a perfusion system in the whole
01:22:51.00 device, if that is necessary. But what is really important,
01:22:54.22 is that we have a light source above, which allows us to actually
01:23:00.07 illuminate the leaves. So we're really maintaining this typical
01:23:06.08 cycle of 16 hours of sunlight, and 8 hours of night. So we're
01:23:11.11 really trying to maintain as close as possible, such a plant in such
01:23:16.19 a microscope. Here's a view from the top, which you can see is completely
01:23:22.16 accessible. Here's the detection system, and here's the illumination
01:23:26.22 lens. And down here, you see the tubes for the perfusion
01:23:32.19 system. So you can just grow one of those plants, and you can see
01:23:37.11 it will happily grow out of the field of view over the course of about
01:23:40.23 15 hours. And what we can do is of course track that if we're
01:23:46.24 interested in the development of the primary root.
01:23:50.14 Or you can also stick to a particular region and look at
01:23:54.00 it at a much higher speed. Here, for example, you could look
01:23:57.16 at microtubule dynamics inside such a plant under various
01:24:02.13 different aspects, or at different cell types. Or, of course, you can
01:24:08.03 look at a specimen along multiple directions. And here is
01:24:11.12 a nice example, as you can see, we're looking at four different
01:24:14.07 or three different planes. And in each of those planes, depending
01:24:18.22 on how thin the specimen actually is, you will notice we will never be
01:24:22.11 able to see all the cells. However, if you take multiple views along
01:24:26.04 different directions, you will be able to merge this information
01:24:29.22 at the very end, and that then gives you a complete view
01:24:33.12 or a more complete view of such a plant, or such a specimen.
01:24:38.07 Now as I said, we're interested in the lateral root growth.
01:24:43.14 And here's one of those examples, you see we're maintaining the
01:24:48.05 specimen in the sunlight. The sun over there indicates when we
01:24:52.04 have a day period, and when the sun's gone away, it's the night
01:24:57.10 period. And as you see, the expression level here of this membrane
01:25:01.24 protein, LTI, changes with this cycle. And a higher intensity
01:25:09.23 actually means a higher expression level. And in red, here you see the
01:25:15.02 nuclei. And we're looking at this for a very long period of time,
01:25:19.24 about 75 hours we have to keep the specimen in here. We're really
01:25:23.19 recording a whole stack of images. You see a projection here again
01:25:27.14 every fifteen minutes. Here's the same data set, now shown
01:25:33.04 in 3-dimensions. We are just rotating it. And as you can see,
01:25:41.08 you really get much, much more detail. You really get a much better feel
01:25:44.12 for the development. We're really only interested in this
01:25:48.11 exit region over here. So we're concentrating on that.
01:25:51.07 And if you look here carefully, you will notice that there are many
01:25:55.04 minute changes in the arrangement of these different cells
01:25:59.16 as the lateral root emerges from the primary root.
01:26:06.11 And this is again a process, as I said, that really takes
01:26:11.02 a long time. And in fact, if you also add the quality control
01:26:16.00 where we just let the lateral root develop for another 36 hours,
01:26:23.11 we're really maintaining the whole specimen, essentially a whole
01:26:27.17 week, inside the microscope. Now, let me give you an idea what the
01:26:36.06 instrument actually looks like. So, here is a previous design that we
01:26:43.20 used, that was used by Philipp Keller during his PhD thesis,
01:26:47.03 and it really shows all the parts that are in the system. So we have the
01:26:51.06 laser up here, we have the cameras down there, we have the specimen
01:26:55.15 chamber over here, and the detection path really is just using the
01:27:00.20 lens here, filter over there, tube lens, and of course the camera.
01:27:06.08 That's it. So it's a really efficient detection system. The illumination
01:27:11.09 part is also relatively efficient, so we just have the laser, the beam
01:27:15.24 expander, an AOTF, and then we have the scanner system here which
01:27:18.24 moves the beam up and down, and then the light is fed
01:27:22.01 into the specimen chamber here from the side. And that's essentially
01:27:27.00 all there is from the optical side. And this was packaged
01:27:32.23 during a project by Patrick Theer and Alexander Atzberger
01:27:39.18 into one of those devices. Again, we still have the camera here
01:27:43.22 in the back, we have the filter wheel, there will be a tube
01:27:46.19 lens here, and then that's the detection lens over here, and we
01:27:50.23 have the specimen chamber down here. The illumination
01:27:53.19 path is a little bit more complicated, but it doesn't really matter
01:27:56.05 here for our purposes. And then the light is fed into the side.
01:27:59.20 And really important is that the stage actually allows us
01:28:04.03 to rotate the specimen, also to move it along three different
01:28:08.03 directions, is actually below. So the top of the specimen chamber
01:28:12.24 is completely available. Which is shown here in these pictures
01:28:18.17 of such a device. And as I pointed out before, this is really
01:28:23.10 important. It's important because it allows us to illuminate
01:28:26.12 the leaf from the top. But it also allows us, for example, to
01:28:32.24 insert mechanical devices that you can use for micromanipulation,
01:28:37.18 and also for microinjection, or also to microinject materials
01:28:46.12 into more complex biological specimens. And of course, you could
01:28:52.11 think of other devices where you illuminate along two
01:28:55.03 directions or where you illuminate and observe along two
01:28:58.22 independent directions. And the big advantage of such systems
01:29:02.02 is certainly that they allow you to record data at a much, much
01:29:05.18 higher rate than if you had just a simple two lens device.
01:29:09.09 What are realistic performances? Well, the recording speed only
01:29:14.22 depends on the camera speed. So if you have a camera that
01:29:18.18 records 200 frames for second, you could probably do that.
01:29:20.12 But you will not be able to move it at this rate, but you can
01:29:24.21 certainly move the specimen at rates up to 100 frames per
01:29:29.01 second. You can use various channels to record the data.
01:29:34.22 We are probably getting close to recording about a stack every second.
01:29:38.19 And of course, you can combine all this with optical tweezers,
01:29:45.00 mechanical tweezers, laser cutters. You can work with fluorescence,
01:29:48.23 scattering contrast, OCT, FCS, FLIM, FRET, all that should be possible.
01:29:54.18 Where does the light, the LSFM, or these light sheet based
01:30:00.23 fluorescence microscopes actually shine? Well,
01:30:03.22 if you compare that, for example, to confocal microscopy,
01:30:07.17 you definitely can work with a much higher dynamic range.
01:30:11.01 You can work with much larger images, so more recent cameras easily provide about 5 million pixels. And of course,
01:30:19.21 you have the possibility to record several directions, which is
01:30:25.14 of course, a big factor that distinguishes such systems
01:30:28.13 from others. So let me just summarize what I've showed
01:30:31.09 you. Well, an important aspect is certainly that these light sheet
01:30:36.06 based microscopes penetrate large specimens relatively well.
01:30:39.19 That is due to the fact that the numerical aperture of the illumination system
01:30:45.17 is relatively small. We only illuminate a single plane in focus,
01:30:51.11 and therefore, cause much less photobleaching and much less
01:30:54.08 phototoxicity. We have this capability of using multiple views
01:30:59.13 and they can be used to compensate artifacts, as well
01:31:02.00 as providing us with an isotropic resolution of such a specimen.
01:31:06.16 And we can also use technologies such as structured illumination
01:31:11.07 to estimate the background contribution, which in addition,
01:31:14.20 improves the resolution and the contrast. Now the big
01:31:21.03 issue is that light sheet based fluorescence microscopy
01:31:25.04 exposes the specimen to orders of magnitude less energy
01:31:29.11 than wide-field and confocal fluorescence microscopes.
01:31:33.08 We record huge amounts of data, we can record many,
01:31:38.15 many frames per second, with millions of pixels and an
01:31:42.15 excellent dynamic range. And that of course, provides
01:31:46.06 a solid basis for a really modern approach to cell biology,
01:31:50.04 plant biology, developmental biology, so an entirely different way
01:31:55.10 of recording data and recording huge amounts of data.
01:31:59.18 And of course, we can combine it with any other technique
01:32:03.13 that is currently used in fluorescence microscopy. And I'm quite sure
01:32:08.20 we'll also invent completely new ones in the future.
01:32:13.00 And of course, all this would not be possible if we didn't
01:32:17.00 have excellent funding. So EMBL supported this for many years,
01:32:21.04 but of course we now have funding from the DFG, the Goethe
01:32:24.08 University, the BMBF. And all this money would be useless
01:32:30.01 if we didn't have excellent collaborators. I'm really grateful
01:32:32.18 that I was allowed to collaborate with Johann Wittbrodt for many years,
01:32:36.05 and with Michael Knop, and their members of their lab, Jan
01:32:41.04 Ellenberg, Eric Karsenti, were very important in the very beginning.
01:32:45.19 And of course, we had many discussions with many other
01:32:47.24 people. And currently, I'm really grateful that we can collaborate
01:32:51.07 with Alexis Maizel from the University of Heidelberg in our
01:32:56.11 work on the plant biology. And with this, I thank you
01:33:00.21 very much for your attention.

This Talk

Speaker: Ernst Stelzer

Audience:

Researcher

Recorded: May 2012

Talk Overview

In this talk about light sheet sectioning, Ernst Stelzer discusses the new technique of light sheet microscopy, also known as selective plane illumination (SPIM). This uses two objectives, one to illuminate the sample and a second to image it, and allows long-term 3D imaging of thick specimens, like developing embryos, with minimal photobleaching and phototoxicity.

Questions

What is the fundamental idea behind light sheet imaging?
Why is phototoxicity and photobleaching reduced?
What are the advantages of doing multiple rotation angles?

Answers

View Answers

Speaker Bio

Ernst Stelzer

Dr. Stelzer trained as a physicist and has used his expertise to develop numerous optical techniques for studying problems in the life sciences. Currently his lab focuses on using light sheet-based fluorescence microscopy to study cells cultured in 3D. Stelzer is Vice-Director of the Buchmann Institute for Molecular Life Sciences and Professor in the Life… Continue Reading